So my VHS conversion technique is: capture; I capture the vhs through RCA to the pinnacle rs-232 looking cable using virtual dub using picvideo codec at q20 with v-dub at 352x480 yuy2, capture filter method a;(4x3 video) lanczos resize to 320x240 capture filter method b;(16x9 video) lanczos resize to 352x240 frame to letterbox/crop size 352x480 in resize, no audio compression. re-compress; run through virtual dub at 1650kbps ffdshow codec MPEG-4 XVID one pass average bit rate, H.263 default settings, null transform filter if any minor edge crop cleanup is needed, deinterlace-blend, sharpen by 15, Lanczos3 resize 640x480 (4x3) with no framing options set(is already letterboxed if 16x9)(will i need to use "ZOOM" feature on HDTV when playing back if letterboxed 16x9?) aud recompress - lameMp3 128-190kbps CBR I realize now that i have a good file to store on the hard drive, its still not best to convert to dvd since the quality should be a bit higher. So i need to refine my process since im pretty much NOT keeping these files on my hard drive but further adding loss to them by converting to DVD. I use convertx to avi to convert them to dvd's so at what part in this process should i part and do something different? Should i use a different vid recompress codec and at higher kbps (4000?), and leave the audio at pcm?
Actually i use huffYUV. right now i just noticed that ffdshow package has the huffyuv codec listed. im testing out FFV1 in the yv12 colorspace with type VLC coder. I remember doing some tests using picvideo-mjpeg2 in the past for capture, which yielded good results. It just takes a fast processor. But i think huffyuv capture then resampled into MPEG4 is the best way to go for image quality vs. space.
should i just keep replying to my post until i get an opinion on this?? ok, i dont know why i thought it worked but virtual dub stated i needed something i ddint have when loading a capture that was captured using ffdshow huffyuv codec, wasnt intersted in figuring it out. So i downloaded the huffyuv 2.2.0 codec today. I got confused on the order to do things. i would like capture vhs, so i choose the 352x480 format to pick up the interlaced material. when i go to resample the interlated material, what order shoudl i do the three of the following deinterlace resize null transform (if crop is needed) if i deinterlace-blend the 352x480 that material still comes out 352x480 , so should i resize before or after the deinterlace. the way im thinking is deinterlace first, then filter(shaprpen or whatever), then resize. and is there a benefit for moving up the picture size to 640x480 after the material is deinterlaced, because it seems to be so much clearer than 352x240.
is there somebody that can help me verify that i am on the right track. i have refined it to the following, and have a VERY nice looking image on DVD now capture at 704x480 huffyuv, resample at xvid-4000kbps, deinterlace blend, crop/resize bicubic 1.00 then letterbox to 720x480, sharpen, audio compression set to lamemp3 128kbps. file is now ready to run through convert-x-to-dvd.
